Active-site models for the nickel-iron hydrogenases: Effects of ligands on reactivity and catalytic properties

Article abstract of DOI:10.1021/ic2012759

Described are new derivatives of the type [HNiFe(SR) ₂(diphosphine)(CO)₃]⁺, which feature a Ni(diphosphine) group linked to a Fe(CO)₃ group by two bridging thiolate ligands. Previous work had described [HNiFe(pdt)(dppe)(CO) ₃]⁺ ([1H]⁺) and its activity as a catalyst for the reduction of protons (J. Am. Chem. Soc.2010, 132, 14877). Work described in this paper focuses on the effects on properties of NiFe model complexes of the diphosphine attached to nickel as well as the dithiolate bridge, 1,3-propanedithiolate (pdt) vs 1,2-ethanedithiolate (edt). A new synthetic route to these Ni-Fe dithiolates is described, involving reaction of Ni(SR) ₂(diphosphine) with FeI₂(CO)₄ followed by in situ reduction with cobaltocene. Evidence is presented that this route proceeds via a metastable μ-iodo derivative. Attempted isolation of such species led to the crystallization of NiFe(Me₂pdt)(dppe)I₂, which features tetrahedral Fe(II) and square planar Ni(II) centers (H ₂Me₂pdt = 2,2-dimethylpropanedithiol). The new tricarbonyls prepared in this work are NiFe(pdt)(dcpe)(CO)₃ (2, dcpe = 1,2-bis(dicyclohexylphosphino)ethane), NiFe(edt)(dppe)(CO)₃ (3), and NiFe(edt)(dcpe)(CO)₃ (4). Attempted preparation of a phenylthiolate-bridged complex via the FeI₂(CO)₄ + Ni(SPh)₂(dppe) route gave the tetrametallic species [(CO) ₂Fe(SPh)₂Ni(CO)]₂(μ-dppe)₂. Crystallographic analysis of the edt-dcpe compund [2H]BF₄ and the edt-dppe compound [3H]BF₄ verified their close resemblance. Each features pseudo-octahedral Fe and square pyramidal Ni centers. Starting from [3H]BF₄ we prepared the PPh₃ derivative [HNiFe(edt)(dppe)(PPh₃)(CO)₂]BF₄ ([5H]BF ₄), which was obtained as a ～2:1 mixture of unsymmetrical and symmetrical isomers. Acid-base measurements indicate that changing from Ni(dppe) (dppe = Ph₂PCH₂CH₂PPh₂) to Ni(dcpe) decreases the acidity of the cationic hydride complexes by 2.5 pK _a^PhCN units, from ～11 to ～13.5 (previous work showed that substitution at Fe leads to more dramatic effects). The redox potentials are more strongly affected by the change from dppe to dcpe, for example the [2]^0/+ couple occurs at E_1/2 = -820 for [2]^0/+ vs -574 mV (vs Fc^+/0) for [1]^0/+. Changes in the dithiolate do not affect the acidity or the reduction potentials of the hydrides. The acid-independent rate of reduction of CH ₂ClCO₂H by [2H]⁺ is about 50 s^-1 (25 °C), twice that of [1H]⁺. The edt-dppe complex [2H]⁺ proved to be the most active catalyst, with an acid-independent rate of 300 s^-1.

Full text of DOI:10.1021/ic2012759

ARTICLE
pubs.acs.org/IC
Active-Site Models for the NickelꢀIron Hydrogenases: Effects of
Ligands on Reactivity and Catalytic Properties
Maria E. Carroll, Bryan E. Barton, Danielle L. Gray, Amanda E. Mack, and Thomas B. Rauchfuss*
Department of Chemistry, University of Illinois at UrbanaꢀChampaign, Urbana, Illinois 61801, United States
S Supporting Information
b
ABSTRACT: Described are new derivatives of the type [HNi-
Fe(SR)2(diphosphine)(CO)3]+, which feature a Ni(diphosphine)
group linked to a Fe(CO)3 group by two bridging thiolate ligands.
Previous work had described [HNiFe(pdt)(dppe)(CO)3]+
([1H]+) and its activity as a catalyst for the reduction of protons
(J. Am. Chem. Soc. 2010, 132, 14877). Work described in this paper
focuses on the effects on properties of NiFe model complexes of the diphosphine attached to nickel as well as the dithiolate bridge, 1,3-
propanedithiolate (pdt) vs 1,2-ethanedithiolate (edt). A new synthetic route to these NiꢀFe dithiolates is described, involving reaction of
Ni(SR)2(diphosphine) with FeI2(CO)4 followed by in situ reduction with cobaltocene. Evidence is presented that this route proceeds via
a metastable μ-iodo derivative. Attempted isolation of such species led to the crystallization of NiFe(Me2pdt)(dppe)I2, which features
tetrahedral Fe(II) and square planar Ni(II) centers (H2Me2pdt = 2,2-dimethylpropanedithiol). The new tricarbonyls prepared in this
work are NiFe(pdt)(dcpe)(CO)3 (2, dcpe = 1,2-bis(dicyclohexylphosphino)ethane), NiFe(edt)(dppe)(CO)3 (3), and NiFe(edt)-
(dcpe)(CO)3 (4). Attempted preparation of a phenylthiolate-bridged complex via the FeI2(CO)4 + Ni(SPh)2(dppe) route gave the
tetrametallic species [(CO)2Fe(SPh)2Ni(CO)]2(μ-dppe)2. Crystallographic analysis of the edt-dcpe compund [2H]BF4 and the edt-
dppe compound [3H]BF4 verified their close resemblance. Each features pseudo-octahedral Fe and square pyramidal Ni centers. Starting
from [3H]BF4 we prepared the PPh3 derivative [HNiFe(edt)(dppe)(PPh3)(CO)2]BF4 ([5H]BF4), which was obtained as a ∼2:1
mixture of unsymmetrical and symmetrical isomers. Acidꢀbase measurements indicate that changing from Ni(dppe) (dppe =
Ph2PCH2CH2PPh2) to Ni(dcpe) decreases the acidity of the cationic hydride complexes by 2.5 pKaPhCN units, from ∼11 to ∼13.5
(previous work showed that substitution at Fe leads to more dramatic effects). The redox potentials are more strongly affected by the
change from dppe to dcpe, for example the [2]0/+ couple occurs at E1/2 = ꢀ820 for [2]0/+ vs ꢀ574 mV (vs Fc+/0) for [1]0/+. Changes in
the dithiolate do not affect the acidity or the reduction potentials of the hydrides. The acid-independent rate of reduction of CH2ClCO2H
by [2H]+ is about 50 sꢀ1 (25 °C), twice that of [1H]+. The edt-dppe complex [2H]+ proved to be the most active catalyst, with an acid-
independent rate of 300 sꢀ1
.
’ INTRODUCTION
Since the first crystallographic and IR spectroscopic evidence
on their distinctive active sites,1 the [NiFe]-hydrogenases have
inspired many synthetic models.2 Early work replicated the
characteristic Ni(SR)2Fe subunit, and subsequent models in-
corporated CO and CN ligands bound to Fe.3ꢀ5 Some early
Ni(SR)2Fe compounds electrocatalyze the production of hy-
drogen from protons, although no intermediates have been
established.4,6 Recently, we described the first hydride-contain-
ing models that mimic both functional as well as certain structural
features of the active sites of these enzymes (Figure 1).7ꢀ9
Specifically, the hydride complex [HNiFe(pdt)(dppe)(CO)3]+
([1H]+) was found to catalyze hydrogen evolution from acids
at about ꢀ1.3 V vs Fc+/0. The enzyme has also been shown to
operate via hydride intermediates.10 Although originally described
as unstable,11 1 was found to be robust, reversibly protonated, and
the resulting hydride is susceptible to substitution reactions.
With the goal of improving the catalytic properties of [1H]+,
we have previously described changes in the iron subsite, leading
to derivatives of the type [HNiFe(edt)(dppe)(PR3)(CO)2]+,
which also catalyze proton reduction.9 In this report, we probe
Figure 1. Left: active site of the [NiFe]-hydrogenases in the SI-state,
omitting possible protonation of terminal thiolate ligands. Right: func-
tional models for the active site as discussed in this work.
the influence of the dithiolate and the ligands on nickel. Unlike
the active site of the [FeFe]-hydrogenase, the Ni and Fe centers
in the [NiFe]-hydrogenases are linked by cysteinyl thiolates,
but our previous models employed propanedithiolate (pdt), so
variations of the thiolate were targeted. Furthermore, the
terminal ligands on Ni in the protein are alkylthiolates, which
are better donor ligands than arylphosphines.12 To achieve the
Received: June 13, 2011
Published: August 25, 2011
r
2011 American Chemical Society
9554
dx.doi.org/10.1021/ic2012759 Inorg. Chem. 2011, 50, 9554–9563
|

Inorganic Chemistry
ARTICLE
Table 1. Selected IR Data (cmꢀ1, CH2Cl2 Solutions) for NiFe(SR)2(dxpe)(CO)3 and Derivatives
diphosphine, thiolate
[(μ-I)NiFe(SR)2(dxpe)(CO)3]+
NiFe(SR)2(dxpe)(CO)3
[HNiFe(SR)2(dxpe)(CO)3]+ (NMR δhydride
)
dppe, S2C3H6
dppe, S2C2H4
dppe, (SPh)2
dcpe, S2C3H6
dcpe, S2C2H4
2096, 2054, 2023
2028, 1952
2082, 2024 (δ ꢀ3.53)
2084, 2025 (δ ꢀ5.7)
2095, 2046, 2028
2030, 1959
2035, 1970 dimera: 2001, 1942
2082, 2032 (unstable)
2095, 2053, 2021
2014, 1940
2078, 2017 (δ ꢀ3.00)
2080, 2019 (δ ꢀ5.3)
2096, 2075, 2053, 2029
2015, 1940
a “Dimer” refers to derivatives assumed to be [(CO)2Fe(SR)2Ni(CO)]2(μ-diphosphine)2. The monomeric NiFe complex was not observed for dppn.
Similar results were obtained when starting with Ni(SPh)2(dcpe).
Scheme 1. Route to NiFe(xdt)(dxpe)(CO)3 Complexes from FeI2(CO)4
targeted changes in the ligands, we relied on the previously
described9 condensation of Ni(SR)2(diphosphine) with Fe2(CO)9,
but we also developed a new route to such complexes that
condenses ferrous carbonyls and Ni(SR)2(diphosphine) fol-
lowed by reduction. We had reported a related reaction of
Fe(pdt)(CO)2(dppe) with NiCl2(dppe) to afford salts of
[(μ-Cl)NiFe(pdt)(dppe)2(CO)]+.13 Until this work, we had
not succeeded in converting such halide-bridged intermediates
to the corresponding hydrides.
Setting aside the large volume of work on nickel-aminodipho-
sphine based catalysts14 and those featuring NiFeS2 cores discussed
above, several Ni-containing bi- and polymetallic complexes have
been shown to catalyze hydrogen evolution. One family of catalysts
is based on Ru derivatives of NiN2S2 and NiS4 metalloligands
(N2S2 = diaminodithiolates and S4 = dithioetherdithiolates).15,16
Figure 2. IR spectra (CH2Cl2 solutions, from top): NiFe(pdt)(dppe)-
Even oligomeric nickel dithiolates have been shown to catalyze
proton reduction.18 In other NiꢀFe complexes, catalysis was not
observed when the Fe center is high spin.17
(CO)3 + I2 (top), intermediate [(μ-I)NiFe(pdt)(dppe)(CO)3]I from the
reaction of Ni(pdt)(dppe) + FeI2(CO)4, 1, and [1H]BF4 (bottom).
(diphosphine)(CO)3]+ (Scheme 1). IR spectra of these
reaction solutions feature νCO bands resembling those for
the corresponding hydride cations, but are shifted to higher
energies by about 20 cmꢀ1. Spectra were somewhat more compli-
cated than expected, probably owing to sample decomposition.
Treatment of a CH2Cl2 solution of purified 1 with I2 gave a simpler
FT-IR spectrum (Figure 2).
In an effort to stabilize the μ-iodide intermediate, we
investigated the reaction of FeI2(CO)4 with Ni(Me2pdt)-
(dppe), where Me2pdt is 2,2,-dimethyl-1,3-propanedithiolate. This
bulky dithiolate19 was found to react with the iron reagent in the
expected way as judged by IR spectra. The initial tricarbonyl
product degraded, however, upon crystallization to the ferrous
iodide derivative NiFe(Me2pdt)(dppe)I2, which was character-
ized by single crystal X-ray diffraction (Figure 3). As for other
Ni(II)Fe(II) species, the NiꢀFe distance suggests the absence of
metalꢀmetal bonding. Related adducts of ferrous iodide with
two thiourea ligands are known.20 The formation of the ferrous
iodide complex is proposed to occur via attack of the iodide
counteranion at [(μ-I)NiFe(pdt)(dppe)(CO)3]+ concomitant
with dissociation of the CO ligands.
’ RESULTS AND DISCUSSION
Synthesis of [NiFe(SR)2(diphosphine)(CO)3]+ Complexes.
The condensation reaction of Fe2(CO)9 and Ni(dithiolate)-
(diphosphine) provides a reliable route to complexes of the type
NiFe(dithiolate)(diphosphine)(CO)3 but cogenerates substan-
tial amounts of Fe2(dithiolate)(CO)6. An alternative method
that circumvents this side reaction was developed that employs
FeI2(CO)4 as an Fe source. Thus, the reaction of FeI2(CO)4
and Ni(pdt)(dppe) followed by addition of 2 equiv of
Cp2Co afforded 1 in ∼30% yield. Using the Fe2(CO)9 and
the FeI2(CO)4 methods, we were able to generate complexes
where pdt had been changed to edt and dppe was changed to dcpe
(1,2-bis(dicyclohexylphosphino)ethane) (Table 1). Many addi-
tional building blocks were evaluated as described in the Support-
ing Information. For example, we initially attempted to prepare
Fe(CO)3 derivatives of Ni(pdt)(dmpe), but these species were
unstable, hence the use of the bulkier dcpe.
The FeI2(CO)4 method is proposed to proceed via
the intermediacy of the μ-iodo cations [(μ-I)NiFe(SR)2
9555
dx.doi.org/10.1021/ic2012759 |Inorg. Chem. 2011, 50, 9554–9563

Inorganic Chemistry
ARTICLE
(CO)2Fe(pdt)Ni(CO)(μ-dppe).11 The reaction of Ni(SPh)2-
(dppbz) appeared to give a similar a product (νCO = 2006(s),
1945(m) cmꢀ1). In view of the successes described above with
Ni(pdt)(dcpe), we examined the reaction Ni(SPh)2(dcpe) +
FeI2(CO)4. Reduction of the iodo intermediate with Cp2Co gave
NiFe(SPh)2(dcpe)(CO)3 complex as indicated in the IR spectrum;
however, this initial product isomerized rapidly to a species spectro-
scopically analogous to [(CO)2Fe(SPh)2Ni(CO)]2(μ-dppe)2.
Figure 3. Structure of NiFe(Me2pdt)(dppe)I2. Selected distances (Å):
Ni--Fe, 3.057; Ni1 P2 2.1703(12); Ni(1)ꢀP(1), 2.1748(12); Ni(1)ꢀ
S(2), 2.2296(12); Ni(1)ꢀS(1), 2.2445(12); Fe(1)ꢀS(2), 2.3806(12);
Fe(1)ꢀS(1), 2.3841(12); Fe(1)ꢀI(2), 2.6165(7); Fe(1)ꢀI(1),
2.6458(7).
Redox Properties of 2, 3, and 4. As previously reported, 1
oxidizes reversibly at ꢀ0.57 V vs Fc0/+ in benzonitrile solution
with [NBu4]PF6 as electrolyte.9 As a solvent for the neutral
complexes, PhCN is superior to MeCN, but the redox potentials
are assumed to be the same, that is, EPhCN = EMeCN 21
. Complex
Using the ferrous iodide method, we prepared three new
NiFe complexes: NiFe(pdt)(dcpe)(CO)3 (2), NiFe(edt)(dppe)-
(CO)3 (3), and NiFe(edt)(dcpe)(CO)3 (4), which are spectro-
scopically similar to NiFe(pdt)(dppe)(CO)3.11 Furthermore, all
are stereochemically nonrigid as indicated by the observation
of single resonances in the 31P NMR spectra of each complex.
Reminiscent of the observations on 1,9 the 31P{1H} NMR
spectrum of 3 consists of a singlet at room temperature, but at
lower temperatures the singlet decoalesces into two singlets (eq 1,
Figure 4). For the analogous edt-dcpe complex 4, the 31P NMR
signal remains a sharp singlet at ꢀ70 °C.
2 oxidizes reversibly at ꢀ0.82 V vs Fc0/+, 250 mV milder than
the [1]0/+ couple. Similarly, the dcpe-edt complex 4 oxidizes
at ꢀ0.74 V, 300 mV milder than the [3]0/+ couple. Thus, the
basicity of the diphosphine significantly affects the redox poten-
tials, but the identity of the dithiolate bridge has little effect. For
1, 2, 3 and 4, a second irreversible oxidation is observed at more
positive potentials than the first oxidation (Table 2). We assign
all couples to one-electron events, as was verified for 1.9
Characterization of New Tricarbonyl Hydrides. Proton-
ation of the reduced species, 2, 3, and 4, with HBF4 Et2O gave
3
the corresponding hydrides, the new salts being [HNiFe(pdt)-
(dcpe)(CO)3]BF4 ([2H]BF4), [HNiFe(edt)(dppe)(CO)3]BF4
([3H]BF4), and [HNiFe(edt)(dcpe)(CO)3]BF4 ([4H]BF4).
1
These complexes were also characterized by 31P{1H} and H
NMR spectroscopy. Illustrative of the stability of these complexes,
the NMR spectrum of a CD2Cl2 solution of the edt-dppe cation
[3H]+ was found to remain unchanged in the presence of 500
equiv of trifluoroacetic acid after 5 h at room temperature.
Salts [2H]+ and [3H]+ were further characterized by single
crystal X-ray diffraction (Figures 6, 7, Table 3). Overall, the NiFe
cations closely resemble [1H]+.7 In the crystallographic analyses,
the hydride ligands were located and refined. The NiꢀHꢀFe
linkage is unsymmetrical in all cases, with the FeꢀH distance
shorter than the NiꢀH distance by 0.18 and 0.37 Å for [2H]BF4
and [3H]BF4, respectively. At 2.68 Å, the NiꢀFe distance is
elongated in the bulkier dcpe complex [2H]+ vs 2.61 and 2.60 Å
in [1H]+ and [3H]+, respectively.7
[HNiFe(edt)(dppe)(PPh3)(CO)2]BF4. The substituted edt de-
rivative [HNiFe(edt)(dppe)(PPh3)(CO)2]BF4 ([5H]BF4) was
prepared by substitution of one CO in [3H]BF4 with PPh3.
Interestingly, NMR spectra indicate the presence of two isomers
in a ratio of ∼2:1. Each isomer exhibits a doublet-of-triplets in
the hydride region of the 1H NMR spectrum. For [HNiFe(pdt)-
(dppe)(PPh3)(CO)2]BF4, wehadpreviouslydetectedtraceamounts
SPh Derivatives. We attempted to prepare NiFe(SR)2 com-
pounds using the bis(monothiolate) complex Ni(SPh)2(dppe). We
expected that the unidentate thiolate would more accurately mimic
-
the μ-cysteinate ligands in the active site of the [NiFe]-hydroge-
nases. Ni(SPh)2(dppe) was indeed found to react with FeI2(CO)4,
and the product of this reaction underwent reduction by Cp2Co.
The same product was obtained from the reaction of Ni(SPh)2-
(dppe) and Fe2(CO)9. Crystallographic analysis indicated that
this derivative is [(CO)2Fe(SPh)2Ni(CO)]2(μ-dppe)2 (eq 2,
Figure 5). The crystallographic result is corroborated by the
31P NMR spectrum. This spectrum features doublets of doublets
at δ 33.9 and at 49.9 (JPP = 30, 60 Hz), which are consistent
with symmetrically oriented SPh ligands. The FT-IR spectrum
(νCO = 1993, 1938 cmꢀ1) resembles that reported for
9556
dx.doi.org/10.1021/ic2012759 |Inorg. Chem. 2011, 50, 9554–9563

Inorganic Chemistry
ARTICLE
Table 2. Selected Electrochemical Properties (V vs Fc+/0) of
FeINiI Complexesa
compound
E1/2, NiFe0/+
ipa/ipc
Epab, NiFe+/2+
NiFe(pdt)(dppe)(CO)3, 1
NiFe(pdt)(dcpe)(CO)3, 2
NiFe(edt)(dppe)(CO)3, 3
NiFe(edt)(dcpe)(CO)3, 4
a Conditions: scan rate 100 mV/s, benzonitrile solution. Epa = peak
potential for cathodic peak for second oxidation, this couple being
irreversible.
ꢀ0.580
ꢀ0.820
ꢀ0.540
ꢀ0.740
0.93
0.95
0.33
0.80
ꢀ0.036
+0.220
ꢀ0.075
+0.027
b
Figure 4. 31P{1H} NMR spectra of 3 (500 MHz, THF-d8 solution) at
various temperatures. The signal at ∼δ60 corresponds to Ni(edt)-
(dppe), and the signal at ∼δ47 is unassigned.
Figure 6. Molecular structure of the cation [2H]+ in [HNiFe-
(pdt)(dcꢀpe)(CO)3]BF4. Hydrogen (except the hydride ligand) atoms
and BF4 are not shown. Selected distances (Å): NiꢀFe, 2.6843(5);
NiꢀS(1), 2.2274(7); NiꢀS(2), 2.2307(7); FeꢀS(1), 2.3208(8); Feꢀ
S(2), 2.3169(8); NiꢀH, 1.90(2); FeꢀH, 1.53(2).
that the unsym isomer is dynamic on the NMR time scale such
that only one signal is observed for the dppe ligand in each
isomer (Figure 8). When a CD2Cl2 solution of [5H]+ is cooled,
the signal assigned to the dppe on the unsym isomer broadens.
Although not examined explicitly, it is likely that dppe in the
symmetric isomer is also nonrigid (Scheme 2). These observa-
tions underscore the stereochemical flexibility of the square
pyramidal nickel sites in these hydrido complexes. Depro-
tonation of [5H]+ with Et3N gave the neutral complex 5,
which was characterized by IR and 1H and 31P NMR
spectroscopy.
Figure 5. Structure of [(CO)2Fe(SPh)2Ni(CO)]2(μ-dppe)2. Selected
distances (Å): NiꢀFe, 2.4567(3); Fe(1)ꢀP(2), 2.2326(5); Ni(1)ꢀ
P(1), 2.1905(5); Ni(1)ꢀS(1), 2.2633(5); Ni(1)ꢀS(2), 2.2784(5);
Fe(1)ꢀS(1), 2.2833 (5); Fe(1)ꢀS(2) 2.3167(6).
of a second isomer, but the mole fraction was so small that the
nature of this minor component was unclear.9 By comparisons
Redox Properties of [2H]+, [3H]+, [4H]+, and [5H]+. The
reduction of the hydrides was also examined. For the couple
[1H]+/0, EMeCN is ꢀ1.29V vs Fc0/+. The [1H]+/0 couple is
partially reversible, whereas little reversibility was detected in
the couples [2H]+/0, [3H]+/0, [4H]+/0, and [5H]+/0 (Table 3).
Changing the solvent to CH2Cl2 did not substantially improve
1
of the H NMR shifts with those for the pdt derivatives, the
major isomer in [5H]BF4 is unsymmetrical (as is the pre-
dominant isomer of [HNiFe(pdt)(dppe)(PPh3)(CO)2]+),9
with the PPh3 ligand in a basal position (Scheme 2). The
occurrence of about 30% of an isomer with PPh3 in the
apical position apparently reflects the moderate steric profile
of edt vs pdt.
the reversibility of any of the hydrides. In the case of [1H]+/0
,
The sym- and unsym-isomers of [5H]+ do not readily inter-
convert although variable temperature 31P{1H} spectra indicate
varying the scan rate from 100 to 750 mV/s improved the
reversibility from ipa/ipc from 0.26 to 0.8. The irreversibility of
9557
dx.doi.org/10.1021/ic2012759 |Inorg. Chem. 2011, 50, 9554–9563

Inorganic Chemistry
ARTICLE
Scheme 2. Dynamic Equilibria Proposed for Isomers of
[HNiFe(edt)(dppe)(PPh3)(CO)2]+
Figure 7. Molecular structure of the cation in [[HNiFe(edt)-
(dppe)(CO)3]+ ([4H]+). Hydrogen (except the hydride ligand)
ꢀ
atoms and BF4 are not shown. Selected distances (Å): Niꢀ
Fe, 2.5958(6); NiꢀS(1), 2.1969(9); NiꢀS(2), 2.2128(9); Feꢀ
S(1), 2.3152(9); FeꢀS(2), 2.328(1); NiꢀH, 1.84(3); FeꢀH,
1.58(4).
Table 3. Selected Bond Distances (Å) for
[HNiFe(xdt)(dxpe)(CO)3]BF4
NiꢀFe
NiꢀH
FeꢀH
Δ(NiꢀH, FeꢀH)
[1H]BF4
[2H]BF4
[3H]BF4
2.61
2.68
2.59
1.64
1.91
1.84
1.46
1.53
1.58
0.18
0.38
0.26
the [1H]+/0 couple is due to an intermolecular conversion of
[1H]0 into 1, as indicated by the appearance of a reversible
couple at ꢀ0.58 V corresponding to [1]+/0 (Table 2). Similarly,
events for the [2]+/0 couple appear upon repeated redox cycles
for the [2H]+/0 couple. Compared to EMeCN for [1H]+/0 at
ꢀ1.29 V, [2H]+ reduces at ꢀ1.49 V (Table 4). Thus, the basicity
of the Ni center more modestly affects the [1H]+/0 couple than
the [1]+/0 couple.
Electrocatalysis. These experiments employed CH2Cl2
solutions of the hydride salts; MeCN solutions are stable
only for a few minutes at room temperature. As can be
seen in Table 4, the reduction potentials for CH2Cl2 and
MeCN solutions are similar. For each hydride complex,
the ipc was found to increase with the addition of the acid
Figure 8. Variable temperature 31P{1H} NMR (202 MHz, CD2Cl2)
spectra of [HNiFe(edt)(dppe)(PPh3)(CO)2]BF4. The signal at δ 68,
assigned to the phosphorus centers of dppe on the unsymmetrical
isomer, broadens at ꢀ70 °C, indicative of slowed rotation of the dppe
ligand at the Ni center. The signal at δ 45 is assigned to PPh3 on the
symmetrical isomer. Signals at δ 72 and 66 correspond to PPh3 on the
unsymmetrical isomer and dppe on the symmetrical isomer, but further
assignments cannot be made, based on this data.
in the regime where the catalytic rate is independent of
indicating catalytic reduction of protons. For the less basic
1=2
MeCN
ic=ip ¼ ðn=0:446ÞðRTk=FνÞ
ð3Þ
dppe derivatives, we used trifluoroacetic acid (pKa
=
12.65),22 whereas for the more basic catalysts, the weaker
chloroacetic acid (pKaMeCN = 15.3) was used. So long as the
pKa of the acid is below that of the NiFe hydride catalyst, the
nature of the acid has little influence on rates in the plateau
region.
[H+] (eq 3).21 For [2H]+, the observed value of [ic/ip]max of ∼15
corresponds to an acid-independent turnover frequency of ∼50 sꢀ1
,
about twice the rate of the related dppe derivative.9
Some insight into the catalytic mechanism is provided by
the effect of [H+] on ic/ip, where ic is the catalytic current
and ip is the peak current for reduction of the hydride
complex in the absence of additional acid. For small values
of [H+], ic increases linearly, consistent with the depen-
dence of the rate of hydrogen evolution on protonation
of the reduced hydride.21 The relative catalytic activity (k)
of the NiꢀFe hydrides was estimated by the value of ic/ip
Like the propanedithiolate [1H]+, the ethanedithiolate
[3H]+ catalyzes hydrogen evolution from CF3CO2H. Ecat
for [1H]+ and [3H]+ differ by only 140 mV, but [3H]+ is far
more active (Figure 9, 10). Specifically, the observed [ic/ip]max
of 35ꢀ40 indicates a turnover frequency of 240ꢀ310 sꢀ1. This
rate is about 10ꢁ that of [1H]+ and is comparable to some
nickel-diaminodiphosphine catalysts.24 When the catalyst con-
centration was halved (0.25 mM vs 0.5 mM), the rate in the
9558
dx.doi.org/10.1021/ic2012759 |Inorg. Chem. 2011, 50, 9554–9563

Inorganic Chemistry
ARTICLE
a
Table 4. Selected Electrochemical Properties (V vs Fc+/0) of [HNiFe(xdt)(dxpe)(CO)3]BF4 in CH2Cl2
complex
E([HNiFe]+/0) (MeCN)
ipa/ipc (MeCN)
E([HNiFe]+/0) (CH2Cl2)
ipa/ipc (CH2Cl2)
[HNiFe(pdt)(dppe)(CO)3]BF4, [1H]BF4
[HNiFe(pdt(dcpe)(CO)3]BF4, [2H]BF4
ꢀ1.29
ꢀ1.49
ꢀ1.33
ꢀ1.41
ꢀ1.54
0.26
0.11
0
ꢀ1.34
ꢀ1.56
ꢀ1.33
ꢀ1.47
ꢀ1.47
0.26
0.10
0.08
0
[HNiFe(edt)(dppe)(CO)3]BF4, [3H]BF4
[HFeNi(edt)(dcpe)(CO)3]BF4, [4H]BF4
[HNiFe(edt)(dppe)(PPh3)(CO)2]BF4, [5H]BF4
a Conditions: 0.5 M catalyst, scan rate 100 mV/s.
0
0
0
Figure 9. Dependence of normalized catalytic current ic (1.3 V vs Fc+/0
)
for 0.5 mM [3H]+ on [CF3CO2H]. The graph shows data for three
experiments.
Figure 10. Cyclic voltammogram of [3H]+ before (solid red line) and
after addition of varying equiv (denoted on right) of CF3CO2H. A
voltammogram in the absence of [3H]+ at 620 equiv of acid is shown
with dotted red line. Conditions: ∼0.5 mM [3H]+ in CH2Cl2, 0.1 M
[NBu4]PF6, scan rate 0.1 V/s, glassy carbon working electrode (d =
3.0 mm); Ag wire pseudoreference with internal Fc (E1/2 ꢂ 0 V); Pt
counter electrode.
acid-independent regime remained unchanged, verifying that
the rate is independent of catalyst concentration.
We previously reported that substitution of CO with PPh3 in
[1H]+ resulted in a 2-fold increase in the rate of catalysis. When
the analogous substitution is made with the ethanedithiolate
[3H]+, increasing amounts of CH2ClCO2H cause an increase in
the reductive current at ꢀ1.5 V. Graphs of ic/ip vs acid concen-
tration begin to plateau at ∼500 equiv of acid, corresponding to
ic/ip of 25. However, the value of ic/ip decreases with additional
acid, an effect we tentatively attribute to catalyst decomposition
(See Supporting Information). On the basis of these results,
[5H]+ operates at a slower rate than the analogous tricarbonyl
derivative [3H]+. Note, however, that [5H]+ exists as a mixture
of two isomers, in a ratio of ∼2:1, and one of the two isomers is
likely a more active catalyst.
too slow to allow precise determination of the pKa; we note,
however, that the pKa’s of the edt-dcpe and pdt-dcpe hydrides
were similar.
Theoretical potentials for the reduction of these hydrides were
estimated from EMeCN and pKaPhCN. Adapting the approach of
Evans and Artero et al.,23 the potential for catalytic proton
reduction, Ecat, was assigned to the potential corresponding to
half the catalytic current in the acid-independent regime. This
value was approximately 100 mV positive of E([HNiFe]+/0).
We estimate that the errors for the overpotentials in Table 5
are (50 mV.
Overpotential. The overpotential of these catalysts for hydro-
gen evolution is determined by the pKa’s and reduction poten-
tials of the hydrides [HNiFe]+/0. Both the relevant pKa’s and
redox potentials were determined in MeCN (or PhCN) solu-
tions, where overpotential can be meaningfully calculated,23
although catalysis was conducted in CH2Cl2 because the catalysts
are more stable in this solvent. Our measurements indicate that
’ CONCLUSION
The [NiFe]-hydrogenases operate via intermediates with a
(μ-H)NiFe(SR)2 core, and this paper contributes new examples
of such bioinspired catalytic intermediates.9 In this report, we
explored the effect of coligands (on nickel) and the dithiolate on
the catalytic and other chemical properties of dithiolato NiFe
hydrides.
A promising new entry to these complexes proceeds via the
reaction of nickel dithiolates with FeI2(CO)4. This ferrous carbonyl
iodide and the related bromide have been widely used for the
preparation of ferrous carbonyls.26,27 Typical substitution reactions
of FeI2(CO)4 replace one or more CO ligands with 2eꢀ donors to
give complexes of the type FeI2L2(CO)2 or [FeX3(CO)3]ꢀ.27 The
(ECH Cl ꢀ EMeCN) is e70 mV (Table 4). Acidꢀbase titrations
were conducted on PhCN solutions of the hydrides, since the
neutral conjugate bases are poorly soluble in MeCN. Using
aniline (pKaMeCN = 10.7), the pKaPhCN of [1H]+ was shown by
31P NMR spectroscopy to be 10.8 vs 10.7 previously determined
by IR spectroscopy.9 The pKaPhCN for the related edt complex
2
2
[3H]+ is 11.3. In the case of the edt-dcpe complex [4H]+, the
PhCN
pKa
was determined to be 13.6 using 4-methoxypyridine
(pKaMeCN 14.23).25 Deprotonation of the pdt-dcpe hydride was
9559
dx.doi.org/10.1021/ic2012759 |Inorg. Chem. 2011, 50, 9554–9563

Inorganic Chemistry
ARTICLE
Table 5. Parameters for Hydrogen Evolution Catalyzed by NiFe Hydrides in CH2Cl2 Solution (Potentials in V vs Fc0/+
)
a
hydride catalyst, proton source
Ecat
overpotentialb
rate,c s-1
pKa of hydrided
[HNiFe(pdt)(dppe)(CO)3]+ ([1H]+), CF3CO2H
[HNiFe(pdt)(dcpe)(CO)3]+ ([2H]+), CH2ClCO2H
[HNiFe(edt)(dppe)(CO)3]+ ([3H]+), CF3CO2H
[HNiFe(edt)(dcpe)(CO)3]+ ([4H]+), CH2ClCO2H
[HNiFe(edt)Ni(dppe)(PPh3)(CO)2]BF4 ([5H]+), CH2ClCO2H
a Ecat is the potential at half-height in the [H+]-independent regime (CH2Cl2 solutions).23 b Overpotentials calculated from |Ecat ꢀ E(H2/HA)|. For
MeCN solutions, E(H2/H+) = ꢀ0.07 V and E(H2/HA) is more negative, depending on the pKa of HA (see Supporting Information).23 c Rate at
[H+]-independent regime (see Figure 9). d In PhCN solutions, determined using pyridine/pyridinium derivatives as base/acid (see text).
ꢀ1.20
ꢀ1.46
ꢀ1.23
ꢀ1.45
ꢀ1.45
0.50
0.59
0.49
0.59
0.54
20
10.7
∼13.6
11.3
13.6
14.0
50
240ꢀ310
20
60ꢀ120
intermediate [(μ-I)NiFe(SR)2(diphosphine)(CO)3]+ is structurally
analogous to the hydrido cations. We had characterized [(μ-Cl)NiFe-
(pdt)(dppe)2(CO)]+.7,13 Both the μ-hydride and the μ-iodo tricar-
bonyls reduce readily to give the targeted Ni(I)Fe(I) derivatives. It is
curious that the μ-iodo complexes are more labile than the corre-
sponding μ-hydrido cations. We ascribe this difference to the
stabilizing influence of the strong σ-donor hydride which stabilizes
CO ligands on these midvalent complexes.28
It is instructive to compare these [HNiFe(xdt)(CO)3ꢀx
-
(PR3)x(diphosphine)]+ catalysts and other NiFe-based catalysts.
Complexes of the present type are simpler to analyze since the
hydride complexes are directly observable. In contrast, for other
NiꢀFe systems, the catalysts are generated in situ and hydride
intermediates are not observed spectroscopically.6,15,18,29 Be-
cause the pKa’s of our hydrides can be determined, overpotentials
can be determined with useful precision. Most active site models
for the NiFe-hydrogenases use Ni centers bound to diamino-
dithiolate (S2N2) ligands.4,6,15 Such tetradentate ligands probably
constrain Ni to a square-planar geometry,30 although square-
pyramidal or octahedral NiS2N2 centers have been observed for
NiꢀRu system where the fifth and sixth site are occupied by
hydride and aquo ligands.16 In the Ni-diphosphine catalysts, the
coordination sphere at Ni is proposed to alternate between
square-pyramidal (with a bridging hydride), tetrahedral, and,
possibly, square planar geometries (Figure 11).
With respect to altering basicities and reduction potentials,
this work revealed that terminal ligands on nickel are more
influential than changes in the dithiolate. On the other hand,
the nature of the dithiolate more strongly affects catalytic rates.
The pKa’s of [1H]+ vs [2H]+ differ by 2 units in PhCN solution.
The diphosphine dcpe is known to substantially enhance the
basicity of its complexes relative to dppe. Angelici has shown that
Figure 11. Proposed pathway for hydrogen evolution catalyzed by
NiFe(SR)2 complexes.
dcpe also causes the reduction of the hydrides to become more
difficult by 200ꢀ250 mV. Therefore, compounds containing
more basic diphosphines on Ni will necessarily have higher
overpotentials for electrocatalytic reduction of protons. Trends
in the basicities and redox potentials for the new NiFe complexes
suggest that protonation occurs primarily at the Fe center,
whereas reduction is shared over both Ni and Fe.
The new NiFe hydrides, like others of this type,9 are active
electrocatalysts for the production of hydrogen. The best cata-
lysts operate at overpotentials of ∼500 mV at limiting rates near
about 300 sꢀ1. Relative to the pdt derivatives, the edt-derived
catalyst [3H]+ displays faster turnover rates, while maintaining
the same overpotential. Complexes [1H]+ and [3H]+ exhibit
very similar spectroscopic properties as well as similar reduction
potentials and acidities. In the acid-independent regime, the
differing rates of proton reduction thus must arise from intra-
molecular processes. As pointed out by Fraze et al.,21 such
intramolecular processes could include the coupling of two
hydride ligands to form a bound dihydrogen or subsequent
dissociation of H2 from the metal. At 298 K, a 10x increase in
rate requires a change in activation energy of 2.6 kcal.
Fe(CO)3(dcpe) is more basic than Fe(CO)3(dppe) by about 4
pKa
attenuated in the bimetallic system. In contrast, the change of Fe
(CO)3 to Fe(PPh3)(CO)2 leads to a ΔpKa = 3ꢀ4. The differing
effect of substitution at Fe vs Ni suggests that the hydride is more
strongly bonded to Fe.
An approximate thermodynamic analysis suggests that in-
creasing the basicity of the diphosphine more strongly affects
the redox properties of the NiFe complexes than their basicities.
The [NiFe(pdt)(dcpe)(CO)3]0/+ couple is 250 mV more nega-
tive than the [NiFe(pdt)(dppe)(CO)3]0/+ couple in MePhCN
solution (Table 2). A similar effect is seen for the dppe-edt and
dcpe-edt complexes. The finding that the Ni-center more
strongly affects the redox potential (6.88 kcal/mol at 298 K)
than the basicity (2.6 kcal/mol at 298 K) suggests that oxidation
of the Fe(I)Ni(I) complexes is localized at Ni, that is, corre-
sponding to the couple FeINiI/FeINiII. Substitution of dppe for
C2H4Cl2
units.31 Thus, it appears that the dcpe vs dppe effect is
The mechanism of hydrogen evolution by these NiꢀFe
models begins with the reduction of the Fe(II)ꢀNi(II) hydride
(Figure 11). In this respect, there exists close parallels between
9560
dx.doi.org/10.1021/ic2012759 |Inorg. Chem. 2011, 50, 9554–9563

Inorganic Chemistry
ARTICLE
the catalytic cycles for the μ-hydrides of NiꢀFe dithiolates and
Fe(II)ꢀFe(II) dithiolates.32 Neither model system how-
ever follows the pathways likely for the biological catalysts: the
[NiFe]-hydrogenases do not operate via Fe(I) intermediates3,33
and the [FeFe]-hydrogenases do not operate via bridging
hydrides.34 It is however encouraging that the basic design
inspired by biology is rather forgiving despite differences in
geometry at Ni, terminal ligation at Fe and Ni, and the nature of
the μ-thiolates. One possible reason for the high activity may be
the stereochemical nonrigidity of the nickel sites in both NiFe-
(SR)2(dxpe)(PR3)x(CO)3ꢀx and the protonated derivatives
[HNiFe(SR)2(dxpe)(PR3)x(CO)3ꢀx]+.
1H NMR (CDCl3): δ 7.84(d), 7.44 (m) (dppe C6H5), 2.25 (d)
(S(CH2)2C), 2.14 (d) (P(CH2)2P), 0.96 (s) (C(CH3)2). 31P{1H}
NMR (CDCl3): δ 55.9(s).
[(I)NiFe(Me2pdt)(dppe)(CO)3]+ and NiFe(Me2pdt)I2(dppe).
In a 50 mL Schlenk flask were dissolved 0.140 g (0.24 mmol) of
FeI2(CO)4 and 0.100 g (0.24 mmol) of Ni(Me2pdt)(dppe) in 5 mL of
CH2Cl2. After stirring the solution for 5 min, the IR spectrum of the
mixture showed νCO = 2095, 2075, 2053, and 2020 cmꢀ1. To the red
solution, 20 mL of hexanes was added to precipitate a red solid. The IR
spectrum of the red solid contains νCO = 2095, 2053, and 2020 cmꢀ1
,
which is similar to that for the species formed by reaction of 1 with I2.
The red solid was extracted into 5 mL of CH2Cl2, and this solution was
treated with 1 equiv of NaBF4. After stirring for 1 h, the mixture was
filtered through Celite, the filtrate was concentrated to 2 mL, and hexanes
were layered over the solution. After storage at ꢀ20 °C, red and brown
crystals had formed, which were identified as NiI2(dppe) and NiFe-
(Me2pdt)I2(dppe), respectively, by X-ray diffraction.
’ EXPERIMENTAL SECTION
Reactions were typically conducted using Schlenk techniques at room
temperature. Most reagents were purchased from Aldrich and Strem,
and unless otherwise stated, were used as received. Solvents were HPLC-
grade and dried by filtration through activated alumina or distilled under
nitrogen over an appropriate drying agent. Ni(pdt)(dppe),35 Ni(SPh)2-
(dppe),36 Ni(SPh)2(dmpe),16 NiCl2(dppbz),37 NiCl2(dcpe),38 NiCl2-
(dppf),39 NiCl2(dppp),40 and 112 were prepared according to modified
literature procedures. Chromatography was conducted with Siliaflash
[(CO)2Fe(SPh)2Ni(CO)]2(μ-dppe)2. A 100 mL Schlenk flask was
charged 0.264 g (0.392 mmol) of Ni(SPh)2(dppe)41 and 0.142 g (0.390
mmol) of Fe2(CO)9 followed by 8 mL of tetrahydrofuran (THF). The
mixture was stirred for 1 h. The solvent was removed under vacuum, and
the residue was extracted into Et2O. This extract was filtered through a
pad of Celite and then evaporated under vacuum. After rinsing the
residue with 2 ꢁ 10 mL of MeOH and 2 ꢁ 10 mL of hexane, a dark
brown solid was obtained. Yield: 0.100 g (31%). IR (CH2Cl2): νCO
P60 from Silicycle (230ꢀ400 mesh). HBF4 Et2O (Sigma-Aldrich) was
3
supplied as 51ꢀ57% HBF4 in Et2O (6.91ꢀ7.71 M). A 2.0 M solution of
HCl in Et2O was purchased from Sigma-Aldrich. Bu4NPF6 was pur-
chased from GFS Chemicals and was recrystallized multiple times by
extraction into CH2Cl2 followed by precipitation by hexane. 1H NMR
spectra (500 or 400 MHz) are referenced to residual solvent relative to
TMS. 31P{1H} NMR spectra (202 or 161 MHz) were referenced to
external 85% H3PO4. FT-IR spectra were recorded on a Perkin-Elmer
Spectrum 100 FT-IR spectrometer.
1993, 1938 cm .
ꢀ1 31P{1H} NMR (CD2Cl2): δ 33.9 (d), 49.9 (d). Single
crystals were obtained from slow diffusion of hexane into a concentrated
CH2Cl2 solution.
NiFe(pdt)(dcpe)(CO)3 and [HNiFe(pdt)(dcpe)(CO)3]BF4, 2
and [2H]BF4. A 100 mL Schlenk flask was charged 0.350 g (0.85
mmol) of FeI2(CO)4 and 0.500 g (0.85 mmol) of Ni(pdt)(dcpe). The
flask was cooled to ꢀ78 °C. To the cooled mixture of solids was added
10 mL of CH2Cl2, resulting in a dark brown homogeneous solution. A
0.5 mL sample was removed (using an ambient temperature syringe)
and checked by FT-IR spectroscopy to confirm the formation of the
proposed μ-iodo intermediate (see Table 1). The remaining cold
reaction solution was treated with 0.320 g (1.70 mmol) of Cp2Co.
The reaction solution became brown-green, and was allowed to warm to
room temperature. After stirring for 1 h at room temperature, the
reaction mixture was evaporated under reduced pressure, and the dark
brown residue was washed and sonicated with 4 ꢁ 20 mL portions of
MeCN (until the washings were colorless). The brown residue was
extracted into 20 mL of CH2Cl2, and this solution of the product
was filtered through a pad of Celite. The CH2Cl2 solution of the
product was concentrated to dryness under vacuum. Yield: 0.4 g (62%).
NiCl2(dcpe). The published synthesis was followed,38 but no NMR
data had been reported. 31P{1H} NMR (CD2Cl2): δ 81.2(s). 1H NMR
(CD2Cl2): δ 1.2ꢀ2.6 (broad m).
Ni(pdt)(dcpe). A solution of 3.50 g (6.34 mmol) of NiCl2(dcpe) in
30 mL of CH2Cl2 was treated with 0.685 g (6.34 mmol) of 1,3-
propanedithiol followed by a solution of 1.37 g (25.3 mmol) of NaOMe
in 10 mL of MeOH. The mixture was stirred for 15 min, during which
time the color of the reaction solution deepened. Solvent was removed
under reduced pressure, and the residue was extracted into CH2Cl2
(20 mL). The slurry was filtered through a pad of Celite to remove NaCl,
and the dark orange filtrate was evaporated. Yield: 3.61 g (97%).
31P{1H} NMR (CD2Cl2): δ 72.3(s). Anal. Calcd for C29H54NiP2S2
(found): C 59.29 (59.57); H, 9.26 (9.33).
IR(CH2Cl2): νCO = 2015, 1941 cm .
ꢀ1 31P{1H} NMR (CD2Cl2):δ 81(s).
Into a solution of 0.150 g (0.207 mmol) of NiFe(pdt)(dcpe)(CO)3 in
Ni(edt)(dcpe). A solution of 1.12 g (2.00 mmol) of NiCl2(dcpe) in
20 mL of CH2Cl2 was treated with 0.188 g (2.00 mmol) of 1,2-
ethanedithiol followed by 0.405 g (4.00 mmol) of triethylamine. The
mixturewas stirred for30 min, during which timethecolor of thereaction
solutiondeepened. Solvent was removed under reduced pressure, andthe
residue was washed with 3 ꢁ 10 mL portions of MeOH. The remaining
orange solid was recrystallized by the addition of hexanes to a CH2Cl2
solution, and the resulting orange crystals were washed with 10 mL
portions of hexanes and MeOH. Yield: 1.00 g (60%). 31P{1H} NMR
(CD2Cl2): δ 78.4(s). Anal. Calcd for C28H52NiP2S2 CH2Cl2 2C6H14
10 mL of CH2Cl2 was injected 100 μL (0.691 mmol) of HBF4 Et2O. An
3
immediate color change from green-brown to dark orange was observed.
The solution was concentrated under reduced pressure to about 5 mL,
diluted with 30 mL of hexanes, and sonicated for 20 min to obtain a dark
orange solid. The solid was filtered, and the reprecipitation method was
repeated four cycles to afford a fine light orange powder. Yield: 0.09 g
1
(53%). 31P{1H} NMR (CD2Cl2): δ 89.8(s). H NMR (CD2Cl2): δ
ꢀ3.00 (br. s, 1H, hydride). Anal. Calcd for C32H55BF4FeNiO3P2S2
3
3
3
CH2Cl2 (found): C, 44.03 (43.72); H, 6.38 (6.41). ESI-MS: m/z 727
(found): C 59.28 (59.58); H, 9.95 (9.24).
([M]+). IR (CH2Cl2): νCO = 2078, 2017 cmꢀ1
.
Ni(Me2pdt)(dppe). A solution of NiCl2(dppe) (2.65 g, 5.0 mmol)
in 200 mL of CH2Cl2 was treated with 0.79 g (5.8 mmol) of Me2pdtH2,
followed by 1.58 mL (11.6 mmol) of triethylamine causing the orange
solution to become dark red. After stirring for 10 min, the solution was
reduced to half volume, transferred to a separatory funnel, and extracted
into an equal volume of water. The volume of the CH2Cl2 layer was
reduced, diluted with hexanes, and stored at ꢀ20 °C, resulting in
dark red crystals, which were washed with hexanes. Yield: 1.8 g (61%).
NiFe(pdt)(dppe)(CO)3 and [(μ-I)NiFe(pdt)(dppe)(CO)3]I.
The procedure developed for NiFe(pdt)(dcpe)(CO)3 was applied
to the synthesis of 1.9 Yield: 0.22 g (38%). A solution of 30 mg
(0.0426 mmol) of 1 in 5 mL of CH2Cl2 was cooled in a ꢀ78 °C bath
and then treated with 10.8 mg (0.0426 mmol) of solid I2. A 0.5 mL
sample was removed and checked by FT-IR spectroscopy: νCO = 2095,
2055, 2025 cmꢀ1 (see Figure 2).
9561
dx.doi.org/10.1021/ic2012759 |Inorg. Chem. 2011, 50, 9554–9563

Inorganic Chemistry
ARTICLE
NiFe(edt)(dppe)(CO)3 and [HNiFe(edt)(dppe)(CO)3]BF4, 3
and [3H]BF4. A 100-mL Schlenk flask was charged with 0.382 g (0.91
mmol) of FeI2(CO)4 and 0.500 g (0.91 mmol) of Ni(edt)(dppe) and
cooled to ꢀ78 °C. Thirty milliliters of CH2Cl2 was added to the cooled
mixture, resulting in a dark brown homogeneous solution. A 0.5 mL
sample was removed (using an ambient temperature syringe) and
checkedbyFT-IR spectroscopytoconfirm theformationoftheproposed
μ-iodo intermediate (see Table 1). The remaining cold reaction solu-
tion was then treated with a precooled (ꢀ78 °C) solution of 0.344 g
(1.82 mmol) of Cp2Co in 20 mL of CH2Cl2, which was transferred via
cannula. The reaction solution became brown-green and was allowed to
warm to room temperature. After stirring for 20 min at room tempera-
ture, the reaction mixture was evaporated under reduced pressure, and
the dark green residue was washed with 2 ꢁ 30 mL of MeCN. The dark
green residue was extracted into 20 mL of CH2Cl2, and the dark green
product was precipitated by addition of 30 mL of hexanes. Yield: 0.42 g
(71%). 31P{1H} NMR (CD2Cl2): δ 650 (s). IR (CH2Cl2): νCO 2029,
1957 cmꢀ1. The protonation was conducted analogously to the proton-
ation of 1. 1H NMR (CD2Cl2): δ ꢀ5.7 (br s, 1H). 31P{1H} (CD2Cl2): δ
72.7. IR (CH2Cl2): νCO = 2084, 2025 cmꢀ1. Anal. Calcd for
C31H29BF4FeNiO3P2S2 (found): C, 44.59 (44.40, 43.72); H, 3.63
(3.62, 3.5).
CH2Cl2 solutions. Cyclic voltammetry experiments were conducted
using a 20 mL one-compartment glass cell with a tight-fitting Teflon top
using a BAS-100 Electrochemical Analyzer. The working electrode was a
glassy carbon (GC) disk (diameter = 3.00 mm). A silver wire was used as
a quasireference electrode, and the counter electrode was a Pt wire. The
electrolyte was 0.1 M [Bu4N]PF6. Ferrocene (sufficient to give ∼1 mM
solution) was added as an internal reference, and each cyclic voltammo-
gram was referenced to this Fc0/+ couple = 0.00 V. iR compensation was
applied to all measurements using the BAS software. Cell resistance was
determined prior to each scan, and the correction was applied to the
subsequently collected cyclic voltammagram. To account for catalysis by
the GC electrode, control experiments were recorded at the same scan
rate and potential range, and this current was subtracted as background.
During prolonged experiments, additional solvent was added to com-
pensate for evaporative loss. For measurements of ip vs [H+], the total
volume of acid solution (∼20 μL aliquots of 0.5 M acid) was less than 5%
of the sample volume, and this effect was ignored. Between scans, the
solution was purged briefly with N2 and the working GC electrode was
removed and polished. The duration of typical electrochemical titrations
was 30 min.
Crystallography. [HNiFe(pdt)(dcpe)(CO)3]BF4 CH2Cl2. A struc-
3
tural model consisting of the host cation plus a disordered [BF4]ꢀ anion
and a dichloromethane solvate molecule was developed. All BꢀF and
F---F distances in the disordered anion were restrained to be similar
(0.01 and 0.02 esd's respectively). Rigid-bond restraints (esd 0.01) were
imposed on displacement parameters for all disordered sites and similar
displacement amplitudes (esd 0.01) were imposed on disordered sites
overlapping by less than the sum of van der Waals radii. The metal
hydride H atom was found in the difference map and was allowed to
refine isotropically. The remaining H atoms were included as riding
idealized contributors and their U’s were assigned as 1.2ꢁ carrier Ueq.
NiFe(edt)(dcpe)(CO)3 and [HNiFe(edt)(dcpe)(CO)3]BF4, 4
and [4H]BF4. A 100 mL Schlenk flask was charged with 0.62 g
(1.46 mmol) of FeI2(CO)4 and 0.84 g (1.46 mmol) of Ni(edt)(dcpe).
To the mixture was added 20 mL of CH2Cl2 at room temperature
resulting in a brown solution. After stirring for 15 min at room
temperature, the solution was treated with 0.55 g (2.92 mmol) of
Cp2Co, and was allowed to stir for 30 min. The reaction mixture was
evaporated under reduced pressure, and the dark brown solid was
washed with 4 ꢁ 15 mL portions of MeCN (until washings were clear).
The crude product was filtered through a pad of silica gel with CH2Cl2 to
remove unreacted starting Ni(edt)(dcpe). The CH2Cl2 solution was
then concentrated to dryness under vacuum. The product could not be
purified beyond removal of unreacted starting material.31P {1H} NMR
(CD2Cl2): δ 94.1(s). The protonation was conducted analogously to
the protonation of 1. 1H NMR (CD2Cl2): δ ꢀ5.3 (br s, 1H). 31P{1H}
[HNiFe(edt)(dppe)(CO)3]B2OHF6 CH2Cl2. A structural model con-
3
sisting of the host ion pair plus one disordered dichloromethane solvate
molecule was developed. Like bond distances and angles were restrained
to be similar between the two orientations of the disordered dichlor-
omethane solvate molecule (esd 0.01 and 0.02 respectively). Similar
displacement amplitudes (esd 0.01) were imposed on disordered sites
overlapping by less than the sum of van der Waals radii. Additionally, the
displacement parameters were restrained to approximate to isotropic
behavior for all disordered sites. The metal hydride was located in the
difference map. The metal hydride distance was allowed to refine, but
theH atom U’s were assigned as 1.5ꢁ Ueq of thecarrier atom (The carrier
atom was assumed to be Fe1because theFeꢀH distance was shorter than
the NiꢀH distance). The hydroxyl atom on the counterion42 was located
in the difference map. The OꢀH bond distance was allowed to refine, but
the H atom U’s were assigned as 1.5ꢁ Ueq of the carrier O atom. The
hydroxyl H atom was found to be hydrogen bonding to F5 on the
counterion. Remaining H atoms were included as riding idealized
contributors and their U’s were assigned as 1.2ꢁ carrier Ueq.
(CD2Cl2): δ 92.0 (s). IR (CH2Cl2): νCO = 2080, 2019 cmꢀ1
.
[HNiFe(edt)(dppe)(PPh3)(CO)2]BF4 and NiFe(edt)(dppe)-
(PPh3)(CO)2, 5 and [5H]BF4. In a 250 mL round bottomed Schlenk
flask, 0.760 g (0.98 mmol) of [HNiFe(edt)(dppe)(CO)3]BF4 was
dissolved in 80 mL of THF. The solution was treated with 2.5 g
(9.8 mmol) of PPh3. After stirring the solution for 2 h at 40 °C, the
solvent was removed in vacuum, yielding a red oil, which was washed
with 3 ꢁ 40 mL portions of hexanes. The resulting reddish-brown
powder was recrystallized by layering a CH2Cl2 solution with hexanes,
which yielded orange crystals. Yield: 0.80 g (80%). 1H NMR (CD2Cl2):
ꢀ5.13 (dt), ꢀ8.15 (dt). 31P{1H} NMR (CD2Cl2): 71.5 (s, PPh3 unsym
or dppe sym), 67.1 (s, dppe unsym), 66.8 (s, PPh3 unsym or dppe sym),
45.5 (s, PPh3, sym). IR (CH2Cl2): νCO = 2019, 1968 cmꢀ1. Anal. Calcd
[(CO)2Fe(SPh)2Ni(CO)]2(μ-dppe)2. A structural model consisting of
the molecule was developed. This model converged with wR2 = 0.0746
and R1 = 0.0438 for 451 parameters against 8306 data. H atoms were
included as riding idealized contributors. H atom U’s were assigned as
1.2ꢁ carrier Ueq.
for C48H44BF4FeNiO2P3S2 CH2Cl2 (found): C, 53.69 (53.34); H, 4.23
3
(4.48). IR (CH2Cl2): νCO = 2014, 1962 cmꢀ1
.
A solution of 0.265 g (0.26 mmol) of [HNiFe(edt)(dppe)-
(PPh3)(CO)2]BF4 in 20 mL of CH2Cl2 and 5 mL of MeOH was
treated with 0.014 g (0.26 mmol) of NaOMe in 10 mL of MeOH. The
mixture stirred for 4 h, during which time the solution changed color
from orange to green. Solvent was removed under vacuum, and the
resulting green residue was washed with 3 ꢁ 15 mL portions of water
and 3 ꢁ 10 mL portions of MeOH. 31P{1H} NMR (CD2Cl2, 20 °C):
NiFe(pdt)(dppe)I2. A structural model consisting of the host plus 1/2
a disordered hexane solvate molecule per asymmetric unit was devel-
oped; however, positions for the idealized solvate molecules were
poorly determined. This model converged with wR2 = 0.1387 and
R1 = 0.0514 for 366 parameters with 0 restraints against 6177 data.
Since positions for the solvate molecules were poorly determined, a
second structural model was refined with contributions from the solvate
molecules removed from the diffraction data using the bypass proce-
dure in PLATON.43 No positions for the host network differed by more
than two su’s between these two refined models. The electron count
74.9 (br, dppe), 54.7 (m, PPh3), 48.4 (br, dppe). IR (CH2Cl2): νCO
=
1968, 1912 cmꢀ1
.
Electrochemistry. The nickelꢀironhydrides presented in this paper
degrade to uncharacterized products in MeCN solution (t1/2 ∼ 60 min).
For this reason, most electrochemical measurements were conducted in
9562
dx.doi.org/10.1021/ic2012759 |Inorg. Chem. 2011, 50, 9554–9563

Inorganic Chemistry
ARTICLE
from the “squeeze”model converged in good agreement with the number of
solvate molecules predicted by the complete refinement. The “squeeze”
data are reported here. Methyl H atom positions, R-CH3, were optimized by
rotation about R-C bonds with idealized CꢀH, R--H, and H--H distances.
Remaining H atoms were included as riding idealized contributors. Methyl
H atom U’s were assigned as 1.5ꢁ Ueq of the carrier atom; remaining H
atom U’s were assigned as 1.2ꢁ carrier Ueq.
(18) Angamuthu, R.; Bouwman, E. Phys. Chem. Chem. Phys. 2009,
11, 5578–5583.
(19) Singleton, M. L.; Jenkins, R. M.;Klemashevich, C. L.;Darensbourg,
M. Y. C. R. Chim. 2008, 11, 861–874.
(20) Bierbach, U.; Saak, W.; Haase, D.; Pohl, S. Z. Naturforsch. 1990,
B45 (1990), 1945–1945.
(21) Fraze, K.; Wilson, A. D.; Appel, A. M.; Rakowski DuBois, M.;
DuBois, D. L. Organometallics 2007, 26, 3918–3924.
(22) Izutsu, K. Acid-Base Dissociation Constants in Dipolar Aprotic
Solvents; Blackwell Scientific Publications: Oxford, U.K., 1990.
(23) Felton, G. A. N.; Glass, R. S.; Lichtenberger, D. L.; Evans, D. H.
Inorg. Chem. 2006, 45, 9181–9184. Fourmond, V.; Jacques, P.-A.; Fonte-
cave, M.; Artero, V. Inorg. Chem. 2010, 49, 10338–10347. See also Koper,
M. T. M.; Bouwman, E. Angew. Chem., Int. Ed. 2010, 49, 3723–3725.
(24) Kilgore, U. J.; Roberts, J. A. S.; Pool, D. H.; Appel, A. M.;
Stewart, M. P.; Rakowski DuBois, M.; Dougherty, W. G.; Kassel, W. S.;
Bullock, R. M.; DuBois, D. L. J. Am. Chem. Soc. 2011, 133, 5861–5872.
(25) Kaljurand, I.; K€utt, A.; Soov€ali, L.; Rodima, T.; M€aemets, V.;
Leito, I.; Koppel, I. A. J. Org. Chem. 2005, 70, 1019–1028.
(26) Hieber, W.; Bader, G. Ber. Deutsch. Chem. Ges. 1928, 61B,
1717–1722. Manuel, T. A. Inorg. Chem. 1963, 2, 854–858. Cohen, I. A.;
Basolo, F. J. Inorg. Nucl. Chem. 1966, 28, 511–520. Li, B.; Liu, T.; Popescu,
C. V.; Bilko, A.; Darensbourg, M. Y. Inorg. Chem. 2009, 48, 11283–11289.
Moreland, A. C.; Rauchfuss, T. B. Inorg. Chem. 2000, 39, 3029–3036.
Verhagen, J. A. W.; Lutz, M.; Spek, A. L.; Bouwman, E. Eur. J. Inorg. Chem.
2003, 2003, 3968–3974. Jiang, J.; Koch, S. A. Inorg. Chem. 2002,
41, 158–160. Turrell, P. J.; Wright, J. A.; Peck, J. N. T.; Oganesyan, V. S.;
Pickett, C. J. Angew. Chem., Int. Ed. 2010, 49, 7508–7511.
(27) Bellachioma, G.; Cardaci, G.; Macchioni, A.; Venturi, C.;
Zuccaccia, C. J. Organomet. Chem. 2006, 691, 3881–3888. Jiang, J.; Lo,
W.; M€uller, P.; Fine, N. D.; Maruani, M.; Raytman, A. Inorg. Chem.
Commun. 2009, 12, 527–529.
’ ASSOCIATED CONTENT
S
Supporting Information. Crystallographic analyses and
b
additional spectroscopic, electrochemical, and preparative de-
tails. This material is available free of charge via the Internet at
http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: rauchfuz@illinois.edu.
’ ACKNOWLEDGMENT
This research was supported by the NIH. We thank David
Schilter for helpful advice.
’ REFERENCES
(1) Cammack, R.; Frey, M.; Robson, R. Hydrogen as a Fuel: Learning
from Nature; Taylor & Francis: London, U.K., 2001; Ogata, H.; Lubitz,
W.; Higuchi, Y. Dalton Trans. 2009, 7577–7587. Fontecilla-Camps, J. C.;
Amara, P.; Cavazza, C.; Nicolet, Y.; Volbeda, A. Nature 2009, 460,
814–822. Fontecilla-Camps, J. C.; Volbeda, A.; Cavazza, C.; Nicolet, Y.
Chem. Rev. 2007, 107, 4273–4303.
(2) Ohki, Y.; Tatsumi, K. Eur. J. Inorg. Chem. 2010, 2011, 973–985.
He, T.; Tsvetkov, N. P.; Andino, J. G.; Gao, X.-F.; Fullmer, B. C.;
Caulton, K. G. J. Am. Chem. Soc. 2010, 132, 910–911.
(3) Tard, C.; Pickett, C. J. Chem. Rev. 2009, 109, 2245–2274.
(4) Canaguier, S.; Artero, V.; Fontecave, M. Dalton Trans. 2008,
315–325.
(28) Benito-Garagorri, D.; Lagoja, I.; Veiros, L. F.; Kirchner, K. A.
Dalton Trans. 2011, 40, 4778–4792.
(29) Duan, L.; Wang, M.; Li, P.; Wang, N.; Wang, F.; Sun, L. Inorg.
Chim. Acta 2009, 362, 372–376.
(30) Grapperhaus, C. A.; Darensbourg, M. Y. Acc. Chem. Res. 1998, 31,
451–459.
(31) Sowa, J. R., Jr.; Zanotti, V.; Facchin, G.; Angelici, R. J. J. Am.
Chem. Soc. 1992, 114, 160–165.
(32) Ezzaher, S.; Capon, J.-F.; Dumontet, N.; Gloaguen, F.; Pꢀetillon,
F. Y.; Schollhammer, P.; Talarmin, J. J. Electroanal. Chem. 2009, 626,
161–170.
(5) Li, Z.; Ohki, Y.; Tatsumi, K. J. Am. Chem. Soc. 2005, 127,
8950–8951. Jiang, J.; Maruani, M.; Solaimanzadeh, J.; Lo, W.; Koch,
S. A.; Millar, M. Inorg. Chem. 2009, 48, 6359.
(33) Wu, H.; Hall, M. B. C. R. Chim. 2008, 11, 790–804. Siegbahn,
P. E. M.; Tye, J. W.; Hall, M. B. Chem. Rev. 2007, 107, 4414–4435.
(34) Gloaguen, F.; Rauchfuss, T. B. Chem. Soc. Rev. 2009, 38,
100–108.
(6) Perra, A.; Davies, E. S.; Hyde, J. R.; Wang, Q.; McMaster, J.;
Schr€oder, M. Chem. Commun. 2006, 1103–1105.
(7) Barton, B. E.; Whaley, C. M.; Rauchfuss, T. B.; Gray, D. L. J. Am.
Chem. Soc. 2009, 131, 6942–6943.
(35) Schmidt,M.;Hoffmann,G.G.J. Organomet. Chem. 1977,124,C5–C8.
(36) Hayter, R. G.;Humiec, F. S. J. Inorg. Nucl. Chem. 1964, 26, 807–810.
(37) Levason, W.; McAuliffe, C. A. Inorg. Chim. Acta 1974, 11, 33–40.
(38) Angulo, I. M.; Bouwman, E.; van Gorkum, R.; Lok, S. M.; Lutz,
M.; Spek, A. L. J. Mol. Catal. A 2003, 202, 97–106.
(39) Corain, B.; Longato, B.; Favero, G.; Ajꢁo, D.; Pilloni, G.; Russo,
U.; Kreiss, F. R. Inorg. Chim. Acta 1989, 157, 259–266.
(40) Bomfim, J. A. S.; de Souza, F. P.; Filgueiras, C. A. L.; de Sousa,
A. G.; Gambardella, M. T. P. Polyhedron 2003, 22, 1567–1573. Van
Hecke, G. R.; Horrocks, W. D., Jr. Inorg. Chem. 1966, 5, 1968–1974.
(41) Rauchfuss, T. B.; Shu, J. S.; Roundhill, D. M. Inorg. Chem. 1976,
15, 2096–2101.
(8) Barton, B. E. Ph.D. Thesis, University of Illinois at Urbana-
Champaign, Urbana, IL, 2010.
(9) Barton, B. E.; Rauchfuss, T. B. J. Am. Chem. Soc. 2010, 132,
14877–14885.
(10) Lubitz, W.; Reijerse, E.; van Gastel, M. Chem. Rev. 2007, 107,
4331–4365.
(11) Zhu, W.; Marr, A. C.; Wang, Q.; Neese, F.; Spencer, D. J. E.;
Blake, A. J.; Cooke, P. A.; Wilson, C.; Schr€oder, M. Proc. Natl. Acad. Sci.
2005, 102, 18280–18285.
(12) Lever, A. B. P. Inorg. Chem. 1990, 29, 1271–1285.
(13) Whaley, C. M. Ph.D. Thesis, University of Illinois at Urbana-
Champaign, Urbana, IL, 2009.
(14) Rakowski DuBois, M.; DuBois, D. L. Acc. Chem. Res. 2009, 42,
1974–1982.
(15) Canaguier, S.; Field, M.; Oudart, Y.; Pꢀecaut, J.; Fontecave, M.;
Artero, V. Chem. Commun. 2010, 46, 5876–5878 For work on related
Ru-Ni systems, see Vaccaro, L.; Artero, V.; Canaguier, S.; Fontecave, M.;
Field, M. J. Dalton Trans. 2010, 39, 3043–3049.
(16) Ogo, S. Chem. Commun. 2009, 3317–3325.
(17) Ichikawa, K.; Matsumoto, T.; Ogo, S. Dalton Trans. 2009,
4304–4309.
(42) Kuhn, N.; G€ohner, M.; Steimann, M. Z. Anorg. Allg. Chem.
2003, 629, 595–596.
(43) Spek, A. L.; van der Sluis, P. Acta Crystallogr. 1990, A46, 194–201.
’ NOTE ADDED AFTER ASAP PUBLICATION
This paper was published on the Web on August 25, 2011, with
the incorrect artwork for the TOC and Abstract graphics. The
corrected version was reposted on August 29, 2011.
9563
dx.doi.org/10.1021/ic2012759 |Inorg. Chem. 2011, 50, 9554–9563